Research on RFID Indoor Positioning Algorithm Based on GRNN Neural Network
Download as PDF
Qian Qiu, Zhitao Dai
The traditional positioning algorithm based on RSSI (Received Signal Strength Indicator) has some problem such as inaccurate ranging, low positioning accuracy and vulnerability to environmental impact. This is because of occlusion, multipath effect and some other factors in indoor positioning using wireless sensor network technology. To solve this problem, a localization algorithm based on the generalized regression neural network (GRNN) is proposed to avoid the negative effect of the parameter n in the prediction propagation model. The algorithm directly establishes the mapping relationship between the RSSI values received by the reference nodes and their position coordinates in the training stage. In the prediction stage, the RSSI values of the nodes to be located are collected and use the learned GRNN neural network localize the location nodes. The simulation results of MATLAB and RFID show that the location algorithm based on GRNN neural network can provide better location results than the path loss model algorithm and the location algorithm based on BP neural network.
Rfid, Indoor Positioning, Rssi, Path Loss Model, Bp Neural Network, Grnn Neural Network